For the last two years I've been slowly working my way towards a new approach to teaching statistics. As my teaching load eases off I've finally had time to work on my R-based introductory statistics course. Don't get me wrong, I love teaching - it's the reason I work here rather than in the research institute I left to take this job. But when you're teaching several classes a day, it just doesn't leave room for other things. Such as thinking.
Finally, the new course is starting to take shape. Its physical form is limited (and not publicly visible) at present, but its shape is growing stronger in my mind. And finally, due to several colleagues here who have helped me, I have finally broken the log jam which has meant that it hasn't been more than a concept up until now. With Eran's help, I can now finally get data into R (it's taken me two years to do that), and so I have been able to explore and tinker with it, gradually building up material that might be suitable for first year students. Progress is slow, but any progress at all is gratifying. Over the past week I've started to think about other issues, such as:
- What role could StatsBytes have in the writing (as opposed to the delivery) of this course? My original plan was to use StatsBytes as the vehicle to construct the course, but because of the problems I've had with R and because I'm adapting an existing course this hasn't really worked out. I'm still hoping the microchunked uncourse approach can play a role in the development of the new module.
- How do I assess competence in a class of nearly 300 students? MCQ's - so hard to resist as the pressure grows.
- And where, if anywhere, do we go with this after the Year 1 module? Is that where StatsBytes comes in?
I don't have answer to these questions yet...